Special observation in
psychiatric hospitals: a
literature review
Report from the Conflict and Containment
Reduction Research Programme
Duncan Stewart, Institute of Psychiatry
Hülya Bilgin, Istanbul University
Len Bowers, Institute of Psychiatry
November 2010
Section of Mental Health Nursing
Institute of Psychiatry
London SE5 8AF
Background
Acutely disturbed inpatients sometimes try to harm themselves, break ward rules, act
aggressively or attempt to leave hospital. Risk of suicide is particularly high for this
population. Special observation (SO) is one way to keep patients safe in these
situations by providing extra care and attention to the patient until s/he is less
disturbed. It involves assigning an identified person to the care of the ‘at-risk’ patient
for a certain period of time, above the minimum general level of observation required
for all inpatients. UK practice guidelines (Department of Health, 1999) suggest three
levels of special observation: intermittent, where the patient’s location must be
checked at specified intervals; within eyesight, where the patient should be kept
within sight at all times; and within arms length, where patients must be observed in
close proximity at all times. The latter two forms of SO are collectively termed
constant SO. This review examines the incidence, duration, antecedents, outcomes
and temporal ecology of SO, as well as the demographic and clinical characteristics
patients under SO.
Literature search
Electronic searches of the main databases were conducted to locate empirical studies
of special observation in English published between 1960 and 2009. The databases
searched were: PsycInfo, Cochrane, Medline, EMBASE Psychiatry, CINAHL and the
British Nursing Index. Key words utilised were observation, special, constant,
intermittent and one-to-one. Resulting titles and abstracts were then inspected for
relevance. The search did not include theses or other grey literature, nor did it
include policy documents or other analyses. The review excludes studies of children,
adolescents and dementia patients, although occasionally these patients were included
in hospital wide analyses of observation. As the literature accumulated, further
references were obtained by following up citations. The final number of identified
empirical papers was 63.
Methodologies of the studies reviewed
Nine of the studies were retrospective analyses of official incident records and/or
patient notes, while 28 used descriptive data from other sources such as surveys,
interviews and observation recording instruments designed for the study. Eight were
classified as before and after studies. There were four case-control studies, three of
which concerned inpatient suicides. The review also includes 19 studies with
qualitative data describing patient and staff experiences of special observation. Two
studies included more than one type of methodology. The studies were conducted in
various types of ward, ranging from acute wards (n=39), forensic units (n=2),
psychiatric unit in a general hospital (n=1), to a mix of wards (including several
categories; n=14). In seven cases, the type of ward was unclear or unspecified. Most
studies were conducted in the UK (n=41) or USA (n=12). Other countries
represented were Canada (n=6), Australia (n=1), Ireland (n=1), Norway (n=1) and a
mix of countries (n=1).
As has been reported previously (Bowers & Park 2001) a huge range of terms are
used to describe various forms of observation. Studies refer to numbered levels of
observation, constant, special, one-to-one, 15-mins, 30-mins, close, primary,
specialing, continuous, and intermittent. To avoid confusion, this review uses the
Department of Health (1999) guidelines to define SO as either intermittent, where the
patients is checked at specified intervals, or constant, where the patient is kept within
sight or reach at all times. A total of 29 studies concerned constant SO only and three
were about intermittent SO only. Twenty studies included both constant and
intermittent SO and these were distinguished in the results, while 5 included both
types but combined them in the results. In six cases there was insufficient
information to determine which form of SO was used.
Analytic procedure
The aim of the literature review was to establish existing evidence for and against the
working model and assess commonality and links between different conflict and
containment types such as patient profiles, chains of events, patient experiences,
circumstances of use, etc. A structured data extraction tool was created with various
headings including sample, methodology, admission status, age, gender, ethnicity,
ward type, service setting, risk status, time spent on ward, rates of special observation,
antecedents/causes, patients’ views, staff views, etc. Where published papers
provided empirical evidence, this was entered on the tool. The headings of the
resultant matrix have then been summarised for the purposes of this review. A
hierarchy of evidence was established to rate the weight of each study in relation to
the project’s aims. The most weight was given to studies conducted in the UK, on
acute wards and/or PICUs, studies with large samples and to findings replicated
across studies.
Incidence
Very few of the reviewed studies provided data on the rate of SO use, and where rates
were reported there was wide variation in the methods employed to calculate them.
The lack of standardised measures in the literature makes it extremely difficult to
compare findings across studies. To aid comparisons, patient and event based SO
rates were calculated, standardised to patients per 100 beds or admissions per month
(Table 1). No event based rates could be calculated.
Table 1: Standardised rates of SO
Paper
Jones et al (2000)
Kettles et al (2004)
Langenbach et al (1999)
Reid & Long (1993)
Shugar & Rehaluk (1990)
Country
Ward type
SO type
UK
UK
Acute
Mixed
UK
UK
Canada
Acute
Acute
Mixed
Constant
Constant
Intermittent
Constant
Constant
Constant
Patient based rate
100 beds
100 adm
50.0
30.8
41.2
20.4
28.3
73.4
12.6
Standardised rates could be calculated for only 5 studies. A Canadian case control
study reported a rate of 73.4 incidents of constant SO use per 100 beds per month,
recorded over an eight month period (Shugar & Rehaluk 1990). The rate per 100
admissions per month was much lower (13), reflecting the brief admission period for
the majority of patients (the mean was 28 hours). Constant SO per 100 beds per
month were calculable for one further study which showed a rate of 50.0 (Jones et al.
2000). Other studies reported rates of 28 patients per 100 admissions per month (Reid
& Long 1993), 72 patients per 100 admissions per month (all types of SO; (Kettles et
al. 2004) and 78 patients per 100 admissions per month (Langenbach et al. 1999).
One UK study reported a mean of 37 hours of SO per 100 occupied bed days (Bowers
et al. 2007). Two studies reported very different results expressed as average hours of
SO per ward week: 445 hours (Porter, McCann, & McGregor 1998) and 44 (Childs,
Thomas, & Tibbles 1994).
Simple percentages of patients subject to SO (during varying time periods) ranged
from 3% (Tardiff 1981), 4% (Dennis 1998) to 8% (Neilson et al. 1996). A study of
238 patients in three London hospitals reported that 47% received intermittent SO
during the first two weeks of admission and 16% received constant SO (Bowers,
Simpson, & Alexander 2003). Equivalent figures for other countries were 14% and
3% in Italy and 19% and 14% in Greece (Bowers et al. 2005). A US survey of 105
hospitals 93% reported using constant SO and 91% intermittent SO (time frame for
these results was not specified)(Green & Grindel 1996).
As the above figures suggest, constant SO tended to be used less frequently than
intermittent or less intrusive forms of enhanced observation. Only one study could be
found with which to compare standardised rates (Table 1) and this showed a higher
rate for intermittent SO. In another UK study rates were 20 patients per 100
admissions per month for constant observation and 58 per 100 admissions for ‘close’
observation (named nurse to be aware of patient location at all times; (Langenbach,
Junaid, Hodgson-Nwaefulu, Kennedy, Moorhead, & Ruiz 1999).
Variations in observation practices between wards in the same hospital have been
found, with some wards recording twice the total hours of observation than others
(Porter, McCann, & McGregor 1998). This study also found substantial differences
in the use of different levels of observation between consultants with the same ward.
Another study of over a thousand patients found statistically significant differences
between wards in the use of enhanced observation levels (Kettles, Moir, Woods,
Porter, & Sutherland 2004). Variations in use of intermittent observation by
consultant before and after service re-configuration have also been reported (Melvin,
Hall, & Bienek 2005). A further study found marked and unexplained variations in
constant SO between hospitals and between wards within the same hospital (Stewart,
Bowers, & Warburton 2009).
Duration
Few studies reported the duration of SO. The mean duration per episode ranged from
34 hours (Pitula & Cardell 1996), 48 hours (Cardell & Pitula 1999), 72 hours
(Bowers, Gournay, & Duffy 2000), 3.8 days (Reynolds et al. 2005), 96 hours (Jones,
Ward, Wellman, Hall, & Lowe 2000), 126 hours (Dennis 1998), 6.8 days (Shugar &
Rehaluk 1990), to 211 hours (Jones, Lowe, & Ward 2000). In most cases, therefore,
SO lasted for at two days or more. A survey of 105 US psychiatric hospitals found
that two thirds reported that patients remained on constant observation for 2 days or
less (Green & Grindel 1996). Very short periods under SO were recorded by one
study, with 70% of patients receiving it for less than 45 minutes (Lehane & Rees
1996).
The minimum reported period was much shorter than this, ranging from 2 hours
(Shugar & Rehaluk 1990), 3 hours (Bowers et al. 2000), 4 hours (Cardell and Pitula,
1999), 12 hours (Dennis, 1998; Jones et al. 2000b), 16 hours (Pitula and Cardell,
1996), 21 hours (Jones et al. 2000a), 2 days (Vrale & Steen 2005).
Some patients were under observation for very long periods indeed. The maximum
reported period ranged from 3.5 days (Pitula and Cardell, 1996), 6 days (Cardell and
Pitula, 1999), 6.8 days (Shugar and Rehaluk, 1990), 10 days (Bowers et al. 2000), 2
weeks (Vrale and Steen, 2005), 372 hours (Dennis, 1998), 408 hours (Jones et al.
200b), 35 days (Green and Grindel, 1996), to 864 hours (Jones et al. 2000a).
Antecedents
Various methodologies were employed to examine the reasons why patients were
placed under SO. Surveys of Trusts and hospitals reveal a wide range of antecedents.
A survey of a random sample of 27 Trusts in England and Wales providing inpatient
services found SO was primarily used to reduce the risk of self-harm and suicide or to
prevent aggressive behaviour or absconding (Bowers et al. 2000). Others included
medical conditions requiring monitoring, confusion, impulsiveness, withdrawal from
drugs and agitation. In the US, hospitals identified safety concerns as the major
criteria for use of constant SO (Green & Grindel 1996). These resulted from
dangerousness to self/others, impulsiveness, inability to make a suicide prevention
contract, psychosis, confused behaviour, the risk of the patient's absconding, risk of
the patient falling. A survey of psychiatric and general hospitals identified the most
common reasons for using constant SO as patients demonstrating aggressive
behaviour toward others, self-harm, suicidal gestures, attempted suicide or acutely
suicidal, patients in mechanical restraints and agitation (Moore et al. 1995). The
study also notes that psychiatric hospitals used constant SO with patients who were
medically ill more frequently than general hospitals.
Descriptive studies of patient data identified actual/risk of self-harm, suicide risk and
aggression as primary reasons for SO. The next most common reason across the
studies was absconding (Childs et al. 1994; Dennis 1998; Langenbach et al. 1999;
Lehane and Rees, 1996; Shugar and Rehaluk, 1990), and in two studies absconding
was the most common reason (Jones et al. 2000a, 200b). When staff were
interviewed about patients observed for absconding risk in one study, risk of violence
or self harm were cited as secondary problems which might occur if the patient left
the ward (Dennis 1997). Other antecedents mentioned included assessment (Jones et
al. 2000b; Langenbach et al. 1999), fire risk (Childs et al. 1994), sexual disinhibition
(Childs et al. 1994), self-neglect (Dennis, 1998), no symptoms (Kettles and Paterson,
2007), first presentation (Kettles and Paterson, 2007; Kettles et al. 2004), safety
considerations (Kettles et al. 2004; Shugar and Rehaluk, 1990), high doses of
medication (Phillips et al. 1977a), medical conditions (Shugar and Rehaluk, 1990),
and to decrease stimulation (Shugar and Rehaluk, 1990). A study of mixed ward
types found levels of observation increased at particular risk periods, such as after
meals for patients with eating disorders and where patients were assessed to be at risk
of not taking their medication (Neilson & Brennan 2001).
Given such an extensive list of antecedents it is disappointing that statistical analysis
of factors associated with special observation was so uncommon. Multivariate
models of official incident data showed statistical associations between constant
observation and absconding, but none with physical aggression, verbal aggression,
self-harm or property damage (Bowers et al. 2007). A logistic regression analysis for
980 patients assessed with a specially devised 12 item instrument found the odds of
patients receiving enhanced observation were increased if patients had mild or severe
suicidal intent, paranoid or delusional beliefs, questionable suicidal intent, severe
agitation, moderate agitation and withdrawn behaviour (Kettles, Moir, Woods, Porter,
& Sutherland 2004). Bowers et al. (2003) found patients categorised (by factor
analysis) as 'self-harmers' were more likely to have received continuous special
observation, as were 'angry absconders' (physical aggression, attempted abscond,
missing without permission) and 'angry-refusers' (aggression, refusing regular and
PRN medication). Categories of patient more likely to receive intermittent
observation were 'abstainers (refusing to eat or drink) and 'medication-ambivalent'
(demanding and refusing prn medication). A study of nurses’ clinical judgement
found that current patient behaviour had more impact than patient history variables on
decisions about which intervention to use, but assaultiveness to others and acting out
were less relevant to observation than to seclusion and restraint (Holzworth & Wills
1999). One analysis found that the 10 most predictive variables (from an initial pool
of 512) in a regression model were: wondering, mental health, physical health,
expansive mood, suicidal ideation, communication, self-esteem/image, interpersonal
relationships, employment and depressed mood (Penny & Frost 1997). With
differences in statistical sophistication and measures used in the analyses it is difficult
to draw any firm conclusions from such a limited number of studies. The results
generally reflect those of the descriptive studies in suggesting a great number of
factors are involved in the decision to place a patient on SO, and not simply whether
patients are at risk of self-harm, suicide or aggression.
The reasons for using special observation are likely to differ by service setting, but
only one study examined this explicitly. A study of forensic and non-forensic units
identified the top six reasons as assault, threat of assault, suicidal intent, self-injury,
threat of self-injury and verbal abuse, but the two forensic units ranked assault first,
while the non-forensic unit ranked suicidal intent first (Whitehead & Mason T. 2006).
Nurses in the forensic units scored patients higher than the non-forensic unit staff in
terms of factors associated with risk of harming themselves or others. There were no
significant differences for psychiatric symptomatology.
The complexities involved in judging whether patients should be place on enhanced
levels of observation are illuminated by some of the qualitative evidence available to
the review. In particular, nurses have indicated that special observation can be used to
manage staff anxieties (Fletcher 1999). Briggs (1974) suggests that staff sometimes
project their own anxiety as that of the patient. Staff identified more circumstances
where special observation is needed than patients were able to, to the extent that use
of the procedure had become a ritual, rather than based on the needs of individual
patients. Anecdotal evidence from one study indicated that nurses felt that special
observation did not foster therapeutic relationships with patients and tended to be used
ritualistically (Moran 1979).
Safety was a predominant theme. Nurses view patient safety as the main concern and
use suicidal indicators such as suicidal ideation/thoughts/plan, depressed mood,
psychotic features (hallucinations), agitation and withdrawal to gauge this (Cleary et
al. 1999). Among 20 patients recently placed under constant observation all
mentioned that the procedure was used for their own protection (Cardell & Pitula
1999). Interviews with eight psychiatric nurses revealed that special observation was
used to keep self-harming patients safe (O'Donovan 2007). This decision was usually
made at assessment, but the paper also notes that special observation could be used
reactively if a patient self-harmed on the unit. The majority of patients and nurses in
one study stated that SO had been used to prevent self-harm (Ashaye, Ikkos, & Rigby
1997).
Circumstances
Initiation
The order for special observation usually came from a doctor (Aidroos 1986;Childs,
Thomas, & Tibbles 1994;Cleary et al. 1999); Gournay and Bowers, 2000; Janofsky,
2009; Kettles and Paterson, 2007; Langenbach et al. 1999; Neilson and Brennan,
2001; O’Donnovan, 2001; Reynolds et al. 2005; Shugar and Rehaluk, 1990; Vrale
and Steen, 2005). Two studies reported that the decision was usually taken by a
multi-disciplinary team (Clark et al. 1999;Hodgson et al. 1993) and three that
multidisciplinary teams were sometimes involved (Bowers et al. 2000; Childs et al.
1994; Langenbach et al. 1999). A survey of NHS Trusts in England and Wales found
that in 50% of cases the initiation of CO was declared to be by joint medical and
nursing decision and in the other half it could be initiated by qualified nurses
(Bowers, Gournay, & Duffy 2000). Only one other study found a significant
proportion of SO episodes to be initiated by nurses. A survey RSUs and special
hospital staff found that 63% indicated that nurses decide on level of observation
(Kitchiner, Riach, & Robinson 1992). Another study found that junior grades of
nurses and doctors tended to carry out the most observation assessments (Kettles,
Moir, Woods, Porter, & Sutherland 2004). There does appear to be room for nurses
to initiate SO in emergency situations, pending a medical assessment circumstances
(Clark et al. 1999; Janofsky, 2009; Shugar and Rehaluk, 1990; Vrale and Steen,
2005), although one study found that nurses initiated SO in only 5% of cases
(Langenbach et al., 1999). If not setting observation levels, nurses can have some
influence and are sometimes consulted (Bowers, Gournay, & Duffy 2000;Gournay
K. & Bowers 2000;O'Donovan 2007).
Practice
Nurses do not always agree with the doctors’ decision to place a patient under special
observation (Cleary et al. 1999; Dennis, 1997) and once initiated, the procedure is
sometimes subject to modification by nurses. One study of 248 episodes of
observation found that in two-thirds of cases nurses did not follow the doctors’ orders
for close (intermittent) observation, at least in terms of near compliance; none
followed them completely (Aidroos 1986). A similar proportion of patients were
placed on close observation even though doctors had ordered general observation than
the other way round. Anecdotal evidence suggested that the most common reason
was that nurses decided for themselves which level of observation was needed. In
another study, interviews with nursing staff suggested that they often modified the
official special observation policy during the course of the procedure, sometimes with
junior doctors’ agreement (Duffy 1995). Senior nurses may have the authority to
review and adjust observation levels in the absence of a medical
officer/multidisciplinary team (Reynolds, O'Shaughnessy, Walker, & Pereira 2005).
One study found that 48/71 patients were placed on SO without a full assessment
(Childs et al. 1994).
Termination
There was greater consensus that special observation is usually terminated by a
doctor. A study of 57 patients whose observation level changed reported that doctors
appeared to continue making the decision to reduce observation levels, with teams
involved in only 2% of cases (Kettles & Paterson 2007). Another found that SO
always terminated formally by a doctor, but often at the suggestion of nursing staff
(Duffy, 1995). Similarly, most staff interviewed in another study reported that
reducing observations was the responsibility of medical staff, although this could be
in discussion with nurses (Pitula & Cardell 1996). Nurses can request that a doctor
terminates constant observation (Shugar & Rehaluk 1990). However, there was
evidence in one unit of (unexplained) unclear decision making between medical and
nursing staff as to when to start or stop constant observation (Janofsky 2009). Only
one study reported that termination of a joint medical and nursing decision in the
majority of cases (Bowers, Gournay, & Duffy 2000). The reliance on doctors can
have unforseen consequences. Duffy (1995) found that since only ward doctors were
responsible for terminating special observation, this rarely occurred at weekends
because only duty doctors were available. This meant that patients could be kept
under observation for longer than necessary.
Who observes
Single site studies suggest a broad range of staff are involved in conducting SO. For
example, one UK study found that observations were primarily carried out by
healthcare support workers without formal training in observation (Reynolds,
O'Shaughnessy, Walker, & Pereira 2005). In the US, registered nurses, mental health
technicians and lay workers or sitters were found to be responsible for SO (Pitula &
Cardell 1996). A broader perspective is provided by larger surveys of SO policy. A
national survey of Trusts in England and Wales found that existing nursing staff
usually carry out SO, but if required bank and agency staff were employed (Bowers et
al. 2000). Trust policies varied as to whether agency nurses, bank nurses or nursing
assistants should be permitted to carry out all levels of SO. Views were polarised as to
whether student nurses and family members were allowed to carry out SO. In the US,
a similar survey reported that a variety of staff provided SO (Green & Grindel 1996).
In order of frequency, these were mental health technicians, registered nurses, nursing
aides, licensed practical nurses, and sitters or nursing assistants. Another found that
almost all constant special observation in psychiatric hospitals was provided by
nurses, licensed practical nurses, nursing assistants and mental health workers, but a
greater range of people were involved in general hospitals (Moore, Berman, Knight,
& Devine 1995). A survey of suicide and self-harm cases subject to legal action found
that of 23 cases where enhanced observation was used, observation was carried out by
unqualified nursing staff in 6 cases, 5 of which were against official Trust policy
(Gournay K. & Bowers 2000). In one hospital, agency nurses accounted for a third of
the total SO time (Childs, Thomas, & Tibbles 1994). It seems that the low status of
SO activity means that nurses often delegate it to unqualified staff (Duffy, 1995).
Temporal & spatial ecology
Very little information was available on the timing and location of special
observation, but it appears that most episodes of SO occur relatively early on during
an admission. Two studies report that most special observation occurred early in an
admission.
Cardell and Pitula (1999) found that SO was used at the time of
admission for 14/20 suicidal patients. It was implemented a few days following
admission in two cases, and for four SO did not occur until months after admission.
The other study reported that 70 % of SO occurred during the first week of admission,
including 40% during the first day when symptoms were most acute (Shugar and
Rehaluk, 1990). Both these studies also reported on the location of special
observation.
Two of the three hospitals in one study routinely conducted SO in
dayroom settings with a TV turned on which patients found disturbing and stressful
(Cardell & Pitula 1999). In the other, use of SO was highest on the crisis unit and
lowest the specialized unit, with over half of episodes beginning on the evening shift
(Shugar and Rehaluk, 1990).
Outcomes
Suicide and self-harm
A small number of studies reported that patients were under SO at the time of suicide.
For intermittent SO the proportion ranged from 18% (Appleby et al. 1999), 45%
(King et al. 2001) to 62% (Busch, Fawcett, & Jacobs 2003). Rates were lower for
constant SO: 2% (Powell et al. 2000), 3% (Appleby et al. 1999) and 9% (Busch,
Fawcett, & Jacobs 2003). However, a case control study found no association
between SO and suicide (Powell, Geddes, Deeks, & Goldacre 2000). Another found
that constant special observation at any time during the admission was higher among
suicide cases (43%) than controls (5%)(Sharma, Persad, & Kueneman 1998). The
difficulty of making clinical judgements as to when SO should be reduced or
terminated is illustrated by two contrasting case studies (Pauker & Cooper 1990). For
one patient, 3.5 months of continuous observation did not reduce life-threatening
acting out and the consultant felt that it had become counter-therapeutic and decided
to gradually reduce observation levels. The patient and family appraised of the
reasons for this and the risks. The patient improved and was eventually discharged.
In a second case, a suicidal episode for a long term patient resulted in increased
neuroleptic dose and constant observation. After signs of improvement, the level of
observation was reduced to intermittent (every 15 minutes). Improvements continued,
but with no warning sign the patient committed suicide. An audit of 31 cases of
suicide and self-harm found that 17 cases required a nurse to be within sight of the
patient at all times and two were under intermittent SO (Gournay and Bowers, 2000).
In all of the constant observation cases, nursed did adhere to the instruction, and while
out of sight the patient committed the act of self-harm or suicide. Another study
controlled for a wide range of patient, conflict, staff and environmental factors, and
found that constant SO was not associated with levels of self-harm (Bowers et al.
2008). However, greater use of intermittent observation was predictive of lower
counts of self-harm. Official counts of self-harm over time have been found to have
no statistical association with trends in the use of constant SO (Stewart et al. 2009).
Other outcomes
One UK study reported outcomes in terms of a generalised measure of untoward
incidents on the wards. No untoward incidents were recorded for patient on the
highest level of observation (observed by a designated nurse in close proximity at all
times) during the first week of admission (Langenbach, Junaid, Hodgson-Nwaefulu,
Kennedy, Moorhead, & Ruiz 1999). Incidents were recorded for 11% of patients on
the next level of observation (nurse to be aware of patient’s location at all times) and
3% of those under general observation. The same study found no significant
correlation between staff sickness and higher levels of patient observation
(Langenbach, Junaid, Hodgson-Nwaefulu, Kennedy, Moorhead, & Ruiz 1999). In
contrast, a study from Canada found a significant correlation between constant SO
and staff sickness (statistic not specified), with sick days increasing after a peak in
continuous observation (Phillips et al. 1977b). Other factors such as changes in
administrative policy or epidemics among staff were not to be associated with
observation hours (details not provided).
Subjects on SO have been found to remain in hospital for longer than other patients
and have a significantly poorer outcome at discharge (Shugar & Rehaluk 1990).
Evidence for the impact of SO on staff sickness levels is mixed. One study found no
association (Langenbach et al. 1999), while another reported a statistically significant
correlation between hours of continuous observation and sick days per month for a
two year period (Phillips et al. 1997b). An intervention study reported reduced rates
of staff sickness after ceasing to use SO, but the association between these two
measures was not formally measured (Dodds and Bowles, 2001).
SO can be a significant drain on hospital resources. If overtime, bank or agency
sources are used the financial cost can mount quickly, exerting pressure on nursing
budgets which generally have little flexibility (Bowers et al. 2000). Childs et al.
(1994) report that over a three month period, the number of supervisions performed
does not initially appear excessive. However, with nurses working a 37.5 hour week,
a total number of 12.98 whole time equivalents (WTE) were needed purely to
supervise, amounting 1.18 WTE per ward. In the US, the estimated cost per day for
the staff person who provided CO ranged from $64 to $575, with a mean of $254.53
(Green and Grindel, 1996). Another US study found constant SO accounted for up to
20% of the total nursing budget, equivalent to $450, 000 (Moore et al. 1995). Three
quarters of units absorbed the cost for these services. The only UK estimates show
intermittent and constant SO to cost £45 million and £35 million respectively across
all inpatient psychiatric wards in England (Flood, Bowers, & Parkin 2008). Cost
savings from reducing SO use have also been reported (Dodds & Bowles 2001).
Other containment methods
Few studies considered SO in the context of other methods of containment. One case
note study of 200 patients found that use of emergency ECT and restraint was not
related to the decision to use of SO (Phillips et al. 1977a). On the other hand, a study
of nurses’ clinical judgement (based upon hypothetical situations) found that
seclusion or restraint was usually recommended in combination with SO (Holzworth
& Wills 1999). The frequency of recommendations for SO only and observation with
seclusion were very similar and both much higher than other treatment options.
Special observation is sometimes used in combination with oral or IV tranquilisation
(Lehane & Rees 1996). In one study, 62% of patients had additional medication
during constant SO, 35 were given IM medication, 12 ECT and five mechanical
restraint (Shugar & Rehaluk 1990). In another, there was no statistical association
between observation level and use of tranquilisers (Langenbach, Junaid, HodgsonNwaefulu, Kennedy, Moorhead, & Ruiz 1999). A comparison across wards found
that one which had never locked its doors had the lowest rate of intermittent and
constant SO observation and the highest level of trained staff (Melvin, Hall, & Bienek
2005). Finally a study of 381 patients receiving various containment measures,
reported that 16% received constant SO only, 14% constant SO and emergency
medication, 2% constant SO and restraint/seclusion, while 19% received constant SO,
emergency medication and restraint/seclusion (Tardiff, 1981).
Characteristics of patients subject to special observation
Very few studies provided demographic information for patients placed under SO.
Where data were presented there was usually no comparison with non-SO patients.
Age
The mean age for SO patients had been reported as 28 (Jones et al. 2000a) and 35
(Jones et al. 2000b). Twenty is the most common lowest age (Ashaye, Ikkos, & Rigby
1997); Jones et al. 2000a, 200b), while on study reported a minimum age of 21 (Pitula
and Cardell, 1996). The highest age was 41 (Jones et al. 2000a), 47 (Pitula and
Cardell, 1996), 59 (Jones et al. 2000b) and 60 (Ashaye et al 1997). One study found
SO patients to be significantly younger, with the highest number of patients in the 1724 and 25-34 age groups (Tardiff, 1981).
Gender
Two studies found no statistically significant relationship between gender and
observation level (Kettles et al. 2004; Tardiff, 1981), while another found a
significantly greater proportion of women under constant SO (Shugar and Rehaluk,
1990). The remainder of studies with available data were descriptive, but all reported
more women than men under SO (Dennis, 1997; Pitula and Cardell, 1996). Of 71
patients placed on SO in one study, 26 were male and 45 were female, but 37 of the
women were under SO because of self-harm (Childs et al. 1994). Similarly, another
study found twice as many women than men under SO among patients diagnosed with
depression and personality disorders (Phillips et al. 1977a).
Ethnicity
Only two studies reported data on the ethnicity of patients under SO. In one, all but
one of the 13 patients studies was white (Ashaye et al. 1997), while the other found
no significant effect for race (Tardiff, 1981).
Legal
One study found legal status to be the factor most strongly related to observation, with
involuntary patients significantly more likely to be placed on the highest level (patient
observed in close proximity; Langenbach et al. 1999). Compared to a control group
involuntary admissions were four times more common among patients placed under
SO (Shugar and Rehaluk, 1990). Other studies found initiation of SO was usually
accompanied by detention under MHA (Duffy, 1995), or the majority of SO patients
to have been detained under the Mental Health Act (Childs et al. 1994; Dennis, 1997).
Diagnosis
Some studies were restricted to suicidal patients, so are excluded from this section.
A ten year retrospective study found diagnosis was significantly correlated with
placement on constant SO. The most frequent diagnosis was schizophrenia (43%),
followed by depression-neurotic and psychotic (22%) and personality disorders (14%)
(Phillips et al. 1997a). Descriptive studies also identified schizophrenia as the most
common diagnosis if patients under SO (Ashaye et al. 1997; Childs et al 1994;
Dennis, 1997; Jones et al. 2000a). Another study found SO patients to be
significantly more likely to be diagnosed with depression, mental retardation or other
nonpsychotic disorders (Tardiff, 1981). Depression was also a frequently reported
diagnosis in other studies (Childs et al., 1994; Jones et al. 2000b). However, one
study found no difference between SO patients and controls in the proportion of
patients diagnosed with affective psychosis, schizophrenia or personality disorder
(Shugar and Rehaluk, 1990). Phillips et al. (1977a) found a significant interaction
between diagnosis and age and gender: SO patients were more likely to be female
than male, with a diagnosis of schizophrenia or depression. If suffering from
depression, she was likely to have been in the 30-50 age group, but if schizophrenia
then the patient was likely to be in the 15-29 or 35-40 age categories. Finally, patients
with diagnoses of psychotic disorders have been reported as significantly more likely
to feel overly observed (Moorhead et al. 1996). The same study reported a trend for
patients who were not caused discomfort to perceive themselves as observed at a
lower intensity than others.
Length of stay
Two studies reported associations between constant SO and length of stay, but one
found SO patients remained in hospital 1.5 times as long as control subjects (Shugar
and Rehaluk, 1990), while the other found patients in hospital for shorter periods were
more likely to receive emergency measures, including constant SO (Tardiff, 1981).
Other
The only other available data concerned the marital status of patients under SO.
Single patients in one study were found to significantly more likely to be placed on
the highest level of SO than others (35% vs 10%)(Langenbach et al. 1999). The
majority of SO patients were reported to be single in another study (Ashaye et al.
1997).
Staff and patient experiences of special observation
There was more evidence available in the literature about staff experiences and views
of SO than patients, although almost all the studies in this section were qualitative and
small scale.
Staff
There was a consistent theme across studies that nurses found SO stressful to conduct.
In particular, many studies mentioned nurses’ unease about the intrusiveness of the
intervention and the potential for patients to react aggressively to it. However, there
was a general consensus that SO was probably needed for some patients, especially
those at risk of self-harm and suicide. Some studies emphasised the need to for good
communication with patients under observation to alleviate some of the negative
consequences (Cleary et al. 1999; Duffy, 1995; Fletcher, 1999). This interaction with
patients leads some nurses to view their role during observations as positive and
therapeutic (Mackay, Paterson, & Cassells 2005); Vrale and Steen, 2005; (Yonge &
Stewin 1992). The introduction of a new SO policy which placed a greater emphasis
on engagement with patients was welcomed by nursing staff and led to increased
patient satisfaction (Reynolds et al. 2005). However, there was also a sense that SO
tends to be a custodial rather than therapeutic intervention (Briggs, 1974; Phillips et
al. 1977a; Reid and Long, 1993; Westhead et al. 2003)
Some studies reported nurses feeling that SO was used inappropriately in some cases.
This included doubts about the suitability of SO as means of preventing absconding
(Ashaye et al. 1997; Clark et al. 1999), as an unconscious means of alleviating staff
anxiety (Briggs, 1974), and doctors placing patients under enhanced observation in
circumstances where nurses judged this to be unnecessary (Cleary et al. 1999; Neilson
and Brennan, 2001). Another found that nurses sometimes felt pressurised by medical
staff to carry out observation against their judgement (Dennis, 1997). A study of
medium secure units reported nurses observing patients with the negative
consequences to themselves in mind should an incident occur (Mason T., MasonWhitehead, & Thomas 2009).
Staffing issues expressed included having adequate numbers on the ward during
observations (Kitchiner, Riach, & Robinson 1992;Westhead et al. 2003), the need to
consider the gender of the observer (Mason T., Mason-Whitehead, & Thomas
2009)Neilson and Brennan 2001; Yonge and Stewin, 1992), and concerns about
whether nursing assistants and agency staff are suitable to conduct SO (Clark et al.
1999).
Patients
Fewer studies (n=11) reported patient perspectives on SO, but a number of themes
were evident from the available literature. Most studies of patient perspectives of SO
report resentment of the associated intrusion and restrictions of being under SO. On
the other hand, the presence of someone with the patient engendered feelings of safety
(Jones et al. 2000b; Pitula and Cardell, 1996).
The attitudes of staff have a strong impact on the patient experience of SO. A
friendly, communicative and positive attitude is appreciated by patients and lessens
the negative aspects of the procedure (Jones et al. 2000a). Thus interviews with 14
suicidal patients reported feelings of discomfort, anxiety and distress if staff showed a
lack of empathy or acknowledgement (Cardell and Pitula, 1999). Another study also
found that suicidal patients in particular want to be observed by nurses they knew and
talk them (Jones et al. 2000b). Lack of familiarity has been identified as a reason for
negative attitudes among patients towards agency staff who observe them (Dodds &
Bowles 2001). In contrast, one study found positive perceptions of lay workers
because they don't monitor participants constantly during personal hygiene or
toileting and are more likely to engage patients in activities (Pitula and Cardell, 1996).
Interviews with 13 patients who had received constant SO found that 11 patients
benefited from it, nine were satisfied with their interactions with staff during the
procedure (Ashaye et al. 1997). Only a quarter of patients in one study reported that
they had been involved in the decisions made regarding observation levels (Reynolds
et al. 2005).
Changes to SO policy and use
Six papers describe the outcome of changes to ward management and SO policies.
Four were specifically designed to reduce SO use, although two papers report results
from the same hospital which reduced SO use to zero over an 18 month period
(Dodds & Bowles, 2001). Reductions in self-harm, violence, absconding and staff
sickness were also reported and there was a substantial saving in staffing costs. The
introduction of a behavioural checklist designed for use with all patients successfully
reduced staff hours spent on SO and use of seclusion also reduced, but there was a
slight increase in the number of recorded incidents on the ward (Moran 1979). A
similar intervention was used to reduce patients leaving the ward without permission
(Richmond, Dandridge, & Jones 1991). A reconfiguration of a service in Scotland
from five to four larger wards resulted in increased occupancy, fewer total staff hours
and a 60% reduction in constant SO. Finally, an audit following a new SO policy
(accompanied by a staff training course to support it) found it to have been generally
well implemented, but one of the five wards accounted for 12 of the 14 patients
placed under SO (Dennis, 1998).
None of these studies is particularly convincing. The designs are weak and data is not
subject to rigorous statistical analysis. However, as a whole they do indicate that it is
possible to reduce the use of SO. The consequences for patients are less clear with
some studies reporting much more optimistic findings than others. Further evaluation
is required which takes account of other forms of containment which might be
substituted for SO.
Evidence for and against the working model
Most of the studies included in the review were descriptive accounts of the frequency
and circumstances of SO use. Relatively few reported data which could shed light on
the impact of nursing practice on conflict and containment. The studies demonstrate
that the functions of SO are confused, which makes identifying potential for changing
and improving nursing practice much more difficult. For example, one study suggests
that SO can be used as an alternative to seclusion (Lehane and Rees, 1996). In one
sense this represents a positive reduction in the use of restrictive containment
measures, but also suggests that at least some form of containment is still required.
This is echoed by nurses’ view of SO as an unpleasant but necessary intervention.
Alternatively, reduced SO can be couched in terms of a broader attempt to reduce
control oriented interventions and to offer more structure to the organisation of the
ward and to interaction with patients (Dodds and Bowles, 2001). This approach
appears to have decreased the rate of deliberate self-harm, absconding and violent
incidents and thus the need for SO. In other words, conflict and containment were
reduced together although such a link is not explicitly examined in this study. The
implied role of organisational support and technical mastery is made more explicit by
another study. Better training and giving authority to nurses to review and adjust SO
levels reduced levels of its use and improved the experience for patients (Reynolds et
al. 2005).
A difficulty for the working model is that whilst SO is a form of containment, the
opportunity to engage with patients on a one-to-one basis may have therapeutic
benefits. A number of studies reported that positive appreciation and engagement with
patients helped reduce the risk of untoward incidents occurring, but this was delivered
during the course of SO. Thus, a study of suicidal inpatients perceptions found that
positive attitudes from observers engendered hope and self-confidence in patients and
ultimately greater compliance with the team and medication (Cardell and Pitula
1999). Distractions provided by an observer were regarded as highly therapeutic in
terms of reduced suicidal feelings. Patients’ feelings were negative where these
qualities were felt to be lacking. Similarly, SO can be viewed as an opportunity to
engage in therapeutic interaction with patients rather than being a ritualistic task to
perform (Jones et al. 2000b; Vrale and Steen 2005). Thus, the working model is
supported because the patient’s experience of observation is strongly influenced by
the characteristics and behaviours of the observer, but contrary to the model, this
positive appreciation occurs during containment.
Nurses also recognised the potential of SO to benefit patients. Interviews with six
PICU nurses found they regarded observation as more than simply watching patients,
and involved a complex caring role which included the use of approaches to
minimising aggression and violence during the procedure (Mackay et al. 2005).
These nurses, therefore, regard SO as a procedure which can implemented in a highly
skilled manner, and in this sense technical mastery does not necessarily lead to less
containment. Staff also believe that SO can be useful for those deemed to be most at
risk as an opportunity for quality one-to-one time and means of providing a safe
environment (Westhead et al. 2003).
Support for the role of teamwork in the model is unclear because of disagreements
between medical and nursing staff over the appropriate observation level for patients.
In one study, similar numbers of patients were placed under SO when doctors had
ordered general observation as the other way round (Aidroos, 1986). On balance, lack
of teamwork skill did not lead to increased containment as the working model would
predict. Disempowering nurses from decision making may contribute to the
perception that SO is a low-grade activity which can be carried out by unqualified
staff, thus reducing the skilled application of the procedure and the likelihood of
therapeutic effects. Effective teamwork is also called into question when nurses
routinely amend the prescribed level of SO without informing the consultant.
Points the model has missed
Placing a single patient under the constant special observation of a nurse leaves the
rest of the ward under-resourced. In essence, this makes the procedure relatively
expensive to implement. The influence of this over decisions to use SO at the ward
level, or more likely at a wider hospital or Trust level, is unknown. In theory, keeping
service costs down in this way may result in no constant special observation being
used, reserving its use for the very high risk patients, and/or encouraging the use of
temporary or untrained staff to conduct the procedure (which is commonly reported).
Alternatively, high levels of SO on a ward may have a negative influence on other
patients because there may be fewer staff available for their needs. The working
model does not take account of levels of staffing on the ward. A prospective case note
study found that constant SO was less likely to be used when there was more qualified
staff present on the ward (Langenbach et al. 1999). Higher levels of staff sickness
have been reported when the amount of SO increases (Phillips et al. 1977b). Another
study cryptically referred to lower staff numbers reducing the options available in
responding to incidents, but did not elaborate on the implications for SO or other
forms of containment (Lehane and Rees, 1996).
Clinical judgement is a central theme in the SO literature, particularly for nurses
whose role in the decision making process seems to vary so greatly. However, there
is not the evidence base to make sound judgements about which kind of observation is
best for which patients in which circumstances. Statistical models appear to have
limited clinical utility, while some studies show nurses to regard patients’ current
behaviour as the main criterion for initiating SO rather than a past history of harmful
actions or patients’ mental state (Hodgson et al. 1993; Holzworth and Wills 1992).
The two cases described by Pauker & Cooper (1990) illustrate the difficulty of
gauging the risks associated with reducing observation. The authors argue that, “we
must accept that facilitating patient improvement and autonomy will engender a small
but significant unavoidable risk of mortality" (p.490).
Discussion
Summary
Up to half of psychiatric patients may be placed under special observation during their
admission, but rates of use vary greatly between studies. This sometimes reflects
local nursing policies but there is also evidence of idiosyncratic practice. Doctors
seem to initiate most SO but nurses can have a great deal of influence over how
observation is actually conducted, which does not always follow the prescribed form.
SO is regarded as an intervention for patients at risk of suicide or self-harm, but is in
fact used for a range of purposes including absconding and aggressive behaviour. The
effectiveness of SO is uncertain, but current evidence is stronger for intermittent than
for constant SO. Patients placed under SO tend to be young, female, involuntary
admissions with a diagnosis of schizophrenia or depression. The quality of the
available data is limited, so firm conclusions about the characteristics of SO patients
cannot be drawn. Staff find SO difficult to conduct but necessary intervention while
patients often resent the level of intrusion associated with it. It is not clear whether it
is desirable to stop using SO altogether or if efforts to do so have resulted in increased
use of other forms of containment.
Lessons for future research
There is very little empirical evidence to guide nurses and medical staff as to which
types of SO work best. The few studies that are available are inconclusive. Two
recent studies employed statistical analyses to examine the effects of SO. One found
greater use of intermittent SO was associated with fewer incidents of self-harm but
there was no apparent benefit for constant SO (Bowers et al. 2008). The other found
that hours of constant SO was not predictive of self-harm outcomes (Stewart et al.
2009). However, these studies utilised shift and weekly counts of ward events rather
example, SO may have been initiated in response to self-harm incidents, but self-harm
could also have occurred while patients were under SO. Importantly, the lack of
statistical relationship between constant SO and self-harm could also reflect poor
implementation of observation practices, such as miscommunication between
observers and other staff members or unclear decision making as to when to
commence or stop various levels of observation. Patient level outcome studies of SO
are urgently required. These studies must take into account the role of other forms of
containment which might be used in conjunction with or instead of SO when
managing difficult patients.
The review shows that the practice of SO is poorly understood. Much of the available
research is limited to descriptive accounts of constant SO or a generalised elevated
level of observation. The very limited data on usage makes it almost impossible to
make meaningful comparisons across studies and settings. Of particular not is the
scant information on the use of intermittent SO and the rationale for opting for one
time interval over another. More research is needed on the use of intermittent SO
because current evidence indicates that is a more effective (and less expensive)
method than constant SO, and is also used more often. The primary purpose of SO is
assumed to be self-harm and suicide prevention, but the review has identified a range
of reasons for its use including preventing absconding, aggressive behaviour, property
damage, medical conditions, confusion, impulsiveness, withdrawal from drugs and
agitation.
Future research should pay more attention to the decision making processes before
during and after SO. Wide variations in SO practices and the ad-hoc amendments to
prescribed SO are reported in the literature, making it essential that the decisions that
inpatient staff make in determining what form of SO is used for particular problem
behaviours are better understood.
References
Aidroos, N. 1986, "Nurses' responses to doctors' orders for close observation",
Canadian Journal of Psychiatry, vol. 31, no. 9, pp. 831-833.
Appleby, L., Shaw, J., Amos, T., McDonnell, R., Harris, C., McCann, K.,
Kiernan, K., Davies, S., Bickley, H., & Parsons, R. 1999, "Suicide within 12
months of contact with mental health services: national clinical survey", British
Medical Journal, vol. 318, pp. 1235-1239.
Ashaye, O., Ikkos, G., & Rigby, E. 1997, "Study of effects of constant
observation of psychiatric in-patients", Psychiatric Bulletin, vol. 21, pp. 145-147.
Bowers, L., Allan, T., Simpson, A., Nijman, H., & Warren, J. 2007, "Adverse
incidents, patient flow and nursing workforce variables on acute psychiatric
wards: The Tompkins Acute Ward Study", International Journal of Social
Psychiatry, vol. 53, no. 1, pp. 75-84.
Bowers, L., Gournay, K., & Duffy, D. 2000, "Suicide and self-harm in inpatient
psychiatric units: a national survey of observation policies", Journal of Advanced
Nursing, vol. 32, no. 2, pp. 437-444.
Bowers, L., Whittington, R., Nolan, P., Parkin, D., Curtis, S., Bhui, K., Hackney,
D., Allan, T., & Simpson, A. 2008, "The relationship between service ecology,
special observation and self-harm during acute in-patient care: the City-128
study.", British Journal of Psychiatry.
Bowers, L., Douzenis, A., Galeazzi, G., Forghieri, M., Tsopelas, C., Simpson, A.,
& Allan, T. 2005, "Disruptive and dangerous behaviour by patients on acute
psychiatric wards in three European centres", Social Psychiatry and Psychiatric
Epidemiology, vol. 40, pp. 822-828.
Bowers, L. & Park, A. 2001, "Special Observation in the Care of psychiatric
Inpatients: A Literature Review", Issues in Mental Health Nursing, vol. 22:769786.
Bowers, L., Simpson, A., & Alexander, J. 2003, "Patient-staff conflict: Results of
a survey on acute psychiatric wards", Social Psychiatry & Psychiatric
Epidemiology., vol. 38, no. 7, pp. 402-408.
Busch, K., Fawcett, J., & Jacobs, D. 2003, "Cinical correlates of inpatient
suicide", Journal of Clinical Psychiatry, vol. 64, pp. 14-19.
Cardell, R. & Pitula, C. R. 1999, "Suicidal inpatients' perceptions of theraputic
and non-theraputic aspects of constant observation", Psychiatric Services, vol.
20(8):1066-1070.
Childs, A., Thomas, B., & Tibbles, P. 1994, "Specialist Needs", Nursing Times,
vol. Jan 19; 90(3) pp32-3.
Clark, N., Kiyimba, F., Bowers, L., Jarrett, M., & McFarlane, L. 1999,
"Absconding: nurses views and reactions", Journal of Psychiatric and Mental
Health Nursing., vol. 6, no. 3, pp. 219-224.
Cleary, M., Horsfall, J., Jordan, R., Horsfall, J., Mazoudier, P., & Delaney, J.
1999, "Suicidal patients and special observation", Journal of Psychiatric and
Mental Health Nursing, vol. 6, pp. 461-467.
Dennis, S. 1997, "Close observation: how to improve assessments", Nursing
Times, vol. 93(24):54-56.
Dennis, S. 1998, "Formal observation in the acute in-patient setting:policy,
training and practice", Mental Health Care, vol. 2(1):26-28.
Dodds, P. & Bowles, N. 2001, "Dismantling formal observation and refocusing
nursing activit in acute inpatient psychiatry: a case study", Journal of
Psychiatric and Mental Health Nursing, vol. 8, pp. 183-188.
Duffy, D. 1995, "Out of the shadows; a study of the special observatin of suicidal
psychiatric in-patients", Journal of Advanced Nursing, vol. 21, 944-950.
Fletcher, R. F. 1999, "The process of constant observation: perspectives of staff
and suicidal patients", Journal of Psychiatric and Mental Health Nursing, vol.
6,1: 9-14.
Flood, C., Bowers, L., & Parkin, D. 2008, "Estimating the Costs of Conflict and
Containment on Adult Acute Inpatient Psychiatric Wards", Nursing Economics,
vol. 26, no. 5, pp. 325-324.
Gournay K. & Bowers, L. 2000, "Suicide and self-harm in in-patient psychiatric
units: a study of nursing issues in 31 cases", Journal of Advanced Nursing, vol.
32, no. 1, pp. 124-131.
Green, S. & Grindel, C. G. 1996, "Supervision of suicidal patients in adult
inpatient psychiatric units in general hopsitals", Psychiatric Services, vol. 47, no.
8, pp. 859-863.
Hodgson, C. M., Kennedy, J., Ruiz, P., Langenbach, M., Moorhead, S., &
Junaid, O. 1993, "Who is watching them? A study of the interpretation of the
observation policy in a mental health unit", Psychiatr Bull., vol. 17: 478-479.
Holzworth, R. J. & Wills, C. E. 1999, "Nurses' judgements regarding seclusion
and restrain of psychiatric patients; a social judgement analysis", Research in
Nursing and Health, vol. 22, no. 3, pp. 189-201.
Janofsky, J. S. 2009, "Reducing inpatient suicide risk: using human factors
analysis to improve observation practices", Journal of the American Academy of
Psychiatry & the Law, vol. 37, pp. 15-24.
Jones, J., Lowe, T., & Ward, M. 2000, "Inpatients' experiences of nursing
observation on an acute psychiatruc unit: a pilot study", Mental Health Care, vol.
4, no. 4, pp. 125-129.
Jones, J., Ward, M., Wellman, N., Hall, J., & Lowe, T. 2000, "Psychiatric
Inpatients' Experience of Nursing Observation: A United Kingdom Perspective
New Models of Care for Acute Sector Mental Health Nursing: A need for
engagement not observation", Journal of Psychosocial Nursing, vol. 38, no. 12,
pp. 10-20.
Kettles, A., Moir, E., Woods, P., Porter, S., & Sutherland, E. 2004, "Is there a
relationship between risk assessment and observation level?", Journal of
Psychiatric and Mental Health Nursing, vol. 11, pp. 156-164.
Kettles, A. & Paterson, K. 2007, "Flexible observation: guidelines versus
reality", Journal of Psychaitric and Mental Health Nursing, vol. 14, pp. 373-381.
King, E., Baldwin, D., Sinclair, J., & Campbell, M. 2001, "The Wessex Recent
Inpatient Suicide Study, 2: Case-control study of 59 inpatient suicides", British
Journal of Psychiatry, vol. 178, pp. 537-542.
Kitchiner, N., Riach, G., & Robinson, T. 1992, "Suicide poilicies in secure
environments", Nursing Standard, vol. April22; 6, no. 31, pp. 37-40.
Langenbach, M., Junaid, O., Hodgson-Nwaefulu, C. M., Kennedy, J., Moorhead,
S. R., & Ruiz, P. 1999, "Observation levels in acute psychiatric admissions
Patients' suicide: court rules supervision was inadequate, wrongful death lawsuit
upheld
Psychiatric patient raped: court faults lack of supervision", European Archives
of Psychiatric and Clinical Neuroscience, vol. 249, no. 1, pp. 28-33.
Lehane, M. & Rees, C. 1996, "Alternatives to Seclusion in Psychiatric Care",
British Journal of Nursing, vol. 5, no. 16, pp. 974-979.
Mackay, I., Paterson, B., & Cassells, C. 2005, "Constant or special observations
of inpatients presenting a risk of aggression or violence: nurses' perceptions of
the rules of engagement", Journal of Psychiatric and Mental Health Nursing, vol.
12, pp. 464-471.
Mason T., Mason-Whitehead, E., & Thomas, M. 2009, "Special observations in
forensic psychiatric practice: gender issues of the watchers and the watched",
Journal of Psychaitric and Mental Health Nursing, vol. 16, no. 10, pp. 910-918.
Melvin, M., Hall, P., & Bienek, E. 2005, "Redesigning acute mental health
services: an audit into the quality of inpatient care before and after service
redesign in Grampian", Journal of Psychiatric and Mental Health Nursing, vol.
12, pp. 733-738.
Moore, P., Berman, K., Knight, M., & Devine, J. 1995, "Constant Observation:
Implications for nursing practice", Journal of Psychosocial Nursing, vol. 33, no.
3, pp. 46-50.
Moorhead, S., Langenbach, M., Kennedy, J., Hodgson, C. M., Ruiz, P., &
Junaid, O. 1996, "Observations of the observed: a study of inpatients' perception
of being observed", Irish Journal of Psychological Medicine, vol. 13, pp. 59-61.
Moran, J. C. 1979, "An alternative to Constant Observation: The behavioural
checklist", Perspectives in Psychiatric Care:The Journal for Nurse
Psychotherapists, vol. 17, pp. 114-117.
Neilson, P. & Brennan, W. 2001, "The use of special observations: an audit
within a psychiatric", 8 pp. 147-155.
Neilson, T., Peet, M., Ledsham, R., & Poole, J. 1996, "Does the nursing care plan
help in the management of psychiatric risk?", Journal of Advanced Nursing, vol.
24, pp. 1201-1206.
O'Donovan, A. 2007, "Pragmatism rules: the intervention and prevention
strategies used by psychiatric nurses working with non-suicidal self-harming
individuals", Journal of Psychiatric and Mental Health Nursing, vol. 14, pp. 6471.
Pauker, S. L. & Cooper, A. M. 1990, "Paradoxical patient reactions to
psychiatric life support: clinical and ethical considerations", American Journal
of Psychiatry, vol. 147(4); 488-491.
Penny, W. & Frost, D. P. 1997, "Neural Network Modeling of the Level of
Observation Decision in an Acute Psychiatric Ward", Computers and Biomedical
Research, vol. 30, pp1-17.
Phillips, M., Peacocke, J., Hermanstyne, L., Rosales, A., Rowe, M., Smith, P.,
Steele, C., & Weaver, R. 1977a, "Continuous Observation - Part 1", Canadian
Psychiatr.Assoc J, vol. 22:25-28.
Phillips, M., Peacocke, J., Hermanstyne, L., Rosales, A., Rowe, M., Smith, P.,
Steele, C., & Weaver, R. 1977b, "Continuous Observation Part II, Absenteeism
Amongst Nursing Staff as a Response to the Extra Workload", Canadian
Psychiatr.Assoc J, vol. 22, no. 1, pp. 29-30.
Pitula, C. & Cardell, R. 1996, "Suicidal Inpatients Experience of Constant
Observation", Psychiatric Services, vol. Aug 47(6): 649-651.
Porter, S., McCann, I., & McGregor, A. 1998, "Auditing Suicide Observation
Procedures", Psychiatric Care, vol. 5(1): 17-21.
Powell, J., Geddes, J., Deeks, J., & Goldacre, H. K. 2000, "Suicide in Psychiatric
Hospital In-patients: Risk factors and their predictive power", British Journal of
Psychiatry, vol. 176, pp. 266-272.
Reid, W. & Long, A. 1993, "The role of the nurse providing therapeutic care for
the suicidal patient", Journal of Advanced Nursing, vol. 18:1369-1376.
Reynolds, T., O'Shaughnessy, M., Walker, L., & Pereira, S. 2005, "Safe and
supportive observation in practice: a clinical governance project", Mental health
Practice, vol. 8, no. 8, pp. 13-16.
Richmond, I., Dandridge, L., & Jones, K. 1991, "Changing nursing practice to
prevent elopement", Journal of Nursing Care Quality, vol. 6, no. 1, pp. 73-81.
Sharma, V., Persad, E., & Kueneman, K. 1998, "A closer look at inpatient
suicide", Journal of Affective Disorders, vol. 47, pp. 123-129.
Shugar, G. & Rehaluk, R. 1990, "Continuous observation for psychiatric
inpatients: a critical evaluation", Comprehensive Psychiatry, vol. 1:48-55.
Stewart, D., Bowers, L., & Warburton, F. 2009, "Constant special observation
and self-harm on acute psychiatric wards: a longitudinal analysis", GeneralHospital Psychiatry, vol. 31, pp. 520-530.
Tardiff, K. 1981, "Emergency Control Measures for Psychiatirc Inpatients", The
Journal of Nervous and Mental Disease, vol. 169, no. 10, pp. 614-618.
Vrale, G. & Steen, E. 2005, "The dynamics between structure and flexibility in
constant observation of psychiatric inpatients with suicidal ideation", Journal of
Psychiatric and Mental Health Nursing, vol. 12, pp. 513-518.
Westhead, J., Cobb, S., Boath, E., & Bradley, E. 2003, "Feedback on constant
observations", Mental Health Nursing, vol. 23, no. 1, pp. 4-8.
Whitehead, E. & Mason T. 2006, "Assessment of risk and special observations in
mental health practice: A comparison of forensic and non-forensic settings",
International Journal of Mental Health Nursing no. 15, pp. 235-241.
Yonge, O. & Stewin, L. L. 1992, "What psychiatric nurses say about constant
care", Hospital and Community Psychiatry, vol. 455; 242-247.